Research methods 1 & 2 Flashcards
Seven steps of the scientific method
- Construct a theory
- Generate a hypothesis
- Choose a research method
- Collect data
- Analyze data
- Report the findings
- Revise existing theories
Theory
Collect a general set of ideas about the way the world works
Hypothesis
Form a testable statement guided by theories that makes specific predictions about the relationship between variables
Research Methods
Determine the way in which the hypothesis will be tested (experiments)
Collect data
Take measurements of the outcomes of the test
Analyze Data
Understand the data and discover trends or relationships between the variables
Report the findings
Publish articles in scholarly journals (time consuming)
Revise Theories
Incorporate new information into our understanding of the world
Key benefit of the scientific method
Standardizes the procedure of research and reduces bias
What is the step after collecting data?
Analyze the data to see if it supports or refutes the hypothesis
Anecdotal evidence
Evidence gathered from others’ or ones’ own experience (insufficient to draw scientific conclusions)
- single experience might not be representative of subsequent experiences
-Personal experience might not represent other’s experience
- energy drinks might not affect test performance
Experiment
Scientific tool used to measure the effect of one variable on another. Scientists manipulate the independent variable to observe the effects on the dependent variable
Independent variable
Variable manipulated by the scientist
Dependent variable
Variable being observed by the scientist
Types of groups in an experiment
Experimental and control group (participants should be similar therefore the differences in the dependent variable will most likely be due to the manipulation of the independent variable)
Experimental group
Manipulation of the independent variable
Control group
No manipulation of the independent group
Which group does not receive the experimental treatment
Control group
Within - Participants Design
Manipulating the independent variable within each participant to minimize the effect of participant differences on the dependent variable
Problems of within participants design
Burdens: time consuming and costly
Variability: difficulty of test or improved performance subject to change
Practice effect
An improvement in performance over the course of an experiment as a result of experience, separate from the effect of the independent variable
Between-participants design
One group gets an independent variable manipulated while the other does not
Confounding variable
A variable associated with an independent variable that obscures the effects of the independent variable on the outcome. This variable makes it difficult to draw findings and conclusions from an experiment. Example (systemic differences such as vegetarians in the experimental groups and non vegs in control group)
The confounding variables influence the results, even though they are not the variable being studied
Problems of strict criteria for study groups
Results from very specific groups of participants cannot be generalized to other groups
Population
General people
Sample
Subset of the people selected from the population
Random sample (ideal for generalization)
A subset of people selected randomly so that it best represents the larger population. This ensures that everyone has an equal chance of getting selected
Random assignment
Randomly assigning the participants to either control or experimental group to avoid any biases
Placebo effect
The situation where the individual exhibits a response to a treatment that is not due to its real therapeutic effects
Participant bias
This can influence the results. The results are what they believe. A mock treatment can be given to the control group and then the participants results can be due to their biases and believes
Experimenters bias
Actions made by the experimenter, intentionally or not, that influence the outcome of the experiment
How to know experimenter bias
If it is unknown whether the participants belong to which group
Double blind experiments
Great way to minimize experimenter and participant bias
Neither the experimenter nor the participants know which group they are in
If the experiment is conducted blind then the participants do not know that
Whether they are a member of the control or experimental group
Descriptive statistics
Provides information about data in a glance to give an overall idea of the results (mean, mode, median). These could be Venn diagrams, charts, bar graphs etc.
Histogram
The type of graph used to report the number of times groups of values appear in a data set. X axis is divided into groups of values called bins and the y axis (frequency) measures the number of values that fall into a given bin. Often used for frequency distribution
Normal distribution
A distribution with a characteristic smooth, symmetrical, bell-shaped curve containing a single peak. Normal everyday measures like IQ and test scores are a type of this
Mean (most common)
The average value of a data set (add all the numbers and then divide by the number of items in the that set)
Outliers
Extreme points, distant from others’ in a data set. Mean is susceptible to influence by outliers
Mode
The value that appears most frequently in the set. Tells us about the most typical response when looking at a dataset and only one that can be used for non numerical datasets. For instance, vote for the best ice cream flavour, mode can be used to determine it.
Median
The center value in a data set when the set is arranged numerically. This tells us where the middle of the data set is like the mean but has the advantage that it can’t be pulled in one direction by an outlier.
Central tendencies
Do not sufficiently summarize the data (mean, mode,median)
Measures of variability
Tell us how spread out our data is. Common one is standard deviation
Standard deviation
A measure of the average distance of each data point from the mean. Larger standard deviation means larger spread
Inferential statistics
Statistics that allow us to use results from samples to make inferences about overall, underlying populations. Example is a T test
T test
A statistical test that considers each data point from both groups to calculate the probability that two samples were drawn from the same population. This test produces a P value which is a probability (0-1) indicating the likelihood of this difference being observed even if no “real” difference exists
P value of less than .05 indicates (significant)
Less than 5% probability of obtaining the observed difference if there is no “real” difference
P value greater than .05 (insignificant)
Greater than 5% probability of obtaining the observed difference if there is no “real” difference
Statistical significance
When the difference between 2 groups is due to some true difference between the properties of the two groups and not simply due to random variation
The distribution for the experimental group does not overlap the control group. What does it suggest about the populations of the two groups?
The experimental group belongs to a completely different population
A T test reveals the P value to be .26 and what does this indicate?
There is a 26% chance that this result can be observed even if the hypothesis is incorrect
Type 1 error
Believing a difference when a difference does not exist (false alarm). For example, an ineffective drug believed to be effective
Type 2 error
Failing to see a difference when a difference does exist (miss). Effective drug believed ineffective
Descriptive statistics
Mean, stdev, histogram
Observational studies/research
To study without any unethical manipulations
Correlated/ correlation
A measure of the strength of the relationship between two variables
Correlation coefficient (r)
A number between -1 and 1 indicates both the strength and direction of the correlation. Plus 1 means the variables are perfectly positively correlated (both variables increase). Minus 1 means perfectly negative correlation (One variable increases and the other variable decreases). Approaching 0 means weak correlation and 0 means no relationship between the variables. More towards plus or minus indicates a strong correlation. The direction of the slope has nothing to do with the r value. Correlation does not equal causation. Be wary of the word “cause” in questions as it is false as it is not correlation
Operational definition
Describes the actions or operations that will be made to objectively measure or control a variable
variable
A feature or characteristic that is free to take on (at least two) different values
Level of analysis
Different perspectives emphasize different aspects of research question
